A new paper from Gaeun Son demonstrates how brief category learning of her novel scene wheel stimuli induces category-specific shifts in people's scene representations. Gaeun had participants first learn to separate scene stimuli from a scene wheel (defined within the latent space of a generative adversarial network, see Gaeun's earlier paper for more details) into two categories. After category learning, participants were briefly shown the scenes and had to reconstruct them from working memory. Reconstruction errors showed biases away from the category boundary suggesting the category learning warped the representational scene space. Notably, participants who didn't learn showed no such biases. And, Gaeun replicated the effects in a second study!

This work appears in Psychonomics Bulletin & Review, check it out!.